Tech firms can't wash hands of harmful effects of AI, UN CIO says

Register now

(Bloomberg) --Technology companies have an obligation to address the unintended consequences of advances like artificial intelligence and innovations that enable damaging cyber-attacks, said Atti Riazi, chief information technology officer at the United Nations.

“It’s not the responsibility of the UN and governments,” she said. “It’s the responsibility of technology companies to say we have created this space, it has implications, it has unintended consequences and we have to figure out how to address it. You can’t just create and innovate without thinking.”

For example, AI has the potential to bring many benefits, including the way the UN is using it to better predict trouble spots globally and to monitor human trafficking and deforestation. But Riazi is also worried about predictions it could lead to mass unemployment.

The UN can’t mitigate AI’s negative outcomes or outlaw malicious uses without help, said Riazi, who also serves as assistant secretary general. “Nobody can enforce it," she said of rules that would ban certain uses of AI. “You’ve created this species of technology, internet, with algorithms, which is beyond our understanding and it is growing, and it is growing exponentially.”

For starters, the UN, governments and nongovernmental organizations worldwide have far less money to spend than the private sector, she said. Riazi spoke an in interview last week in Seattle where she was visiting companies including Tableau Software Inc. and organizations such as Microsoft Corp. Co-Founder Paul Allen’s Vulcan. In meetings with technology companies, she said she is raising the need for more work and cooperation.

Cybersecurity is also a “serious issue” that is not being addressed, she said, including concerns there could be a cyber-attack in Syria to worries about aging energy infrastructure.

The UN has signed a contract with Tableau and is using the company’s data-visualization technology to modernize its systems and do a better job of predicting crisis situations in advance rather than responding after the fact, Riazi said. For example, the UN tried using data like monitoring the price of grain to predict possible unrest after the recent elections in Kenya. When voters expect unrest, they stockpile grain, Riazi said.

The organization is also hoping to use a technology called SkyLight developed by Allen’s Vulcan to combat the depletion of fish populations. Vulcan built the system, which uses satellite imagery and data-analysis software to help countries spot and catch unlicensed fishing boats. The UN would like to consider it for issuess like monitoring human and drug trafficking or forest destruction.

“Hopefully we can find a happy medium where we use the data for the good of humankind, but also protect people and people’s data,” she said.

Bloomberg News
Artificial intelligence Machine learning Cyber security Cyber attacks Malware Data security CIO